The efficacy of nicorandil was compared with that of nifedipine in 13 patients with vasospastic angina enrolled in a randomized, placebo-controlled, crossover study. All patients had a coronary spasm during coronary arteriography, either spontaneously or ergometrine-induced. During two consecutive periods of 2 days, patients received active drugs or placebo in a randomized order. Each patient received single oral doses of 30 mg nicorandil, 10 mg nifedipine, and, on 2 days, a placebo. One hour after drug intake, patients underwent an ergometrine test with increasing doses of Methergin (ergometrine) (0.05, 0.10, 0.20, and 0.40 mg every 5 min). After placebo, the tests always were positive, and the ECG changes occurred at the same +/- 1 dose of ergometrine in 10 cases, showing good reproducibility. After nicorandil, the tests were negative in nine patients and positive for a higher or lower dose of ergometrine in three and one patient, respectively (p = 0.0034 vs. placebo). After nifedipine, the tests were negative in five patients and positive for a higher or the same dose of ergometrine in four and four patients, respectively (p = 0.0039 vs. placebo). Nifedipine (10 mg) and nicorandil (30 mg) were equally effective in eight patients; in the remaining five patients, nicorandil had better results (p = 0.06). Nicorandil (30 mg) prevents ergometrine-induced coronary spasm. This compound may be beneficial in patients with vasospastic angina.